Hi scikit-learn community,

I'm experimenting w/ unsupervised Deep Belief Nets (DBN) for dimension
reduction.

Hinton shows good results using a 2000-500-250-125-2 Autoencoder to
cluster a newswire corpus (essentially a neural topic model):

http://www.cs.toronto.edu/%7Ehinton/science.pdf

I'm trying to do something similar using a simple two step process:

1. Layered RBMs, trained w/ Contrastive Divergence (CD), and then

2. Gradient descent backpropagation on the DBN (weights established in step 1)

I have a good handle on the first part (Layered RBM and CD).

I'm using Edwin Chen's wonderfully documented and simple code --
https://github.com/echen/restricted-boltzmann-machines/blob/master/rbm.py

Here's the related tutorial --
http://blog.echen.me/2011/07/18/introduction-to-restricted-boltzmann-machines/

I'm not sure how to implement the second part -- running gradient
descent backpropagation on the weights established by step 1.

Has anyone tried this, or something similar?

I found a library that uses MDP to do something similar --
http://organic.elis.ugent.be/node/270 -- but i'd like to do it all w/
Edwin's code + scikit-learn.


Thanks,
Timmy Wilson

------------------------------------------------------------------------------
All the data continuously generated in your IT infrastructure 
contains a definitive record of customers, application performance, 
security threats, fraudulent activity, and more. Splunk takes this 
data and makes sense of it. IT sense. And common sense.
http://p.sf.net/sfu/splunk-novd2d
_______________________________________________
Scikit-learn-general mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/scikit-learn-general

Reply via email to